FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3110282355> ?p ?o ?g. }

• W3110282355 abstract "Deep learning approaches are nowadays ubiquitously used to tackle computer vision tasks such as semantic segmentation, requiring large datasets and substantial computational power. Continual learning for semantic segmentation (CSS) is an emerging trend that consists in updating an old model by sequentially adding new classes. However, continual learning methods are usually prone to catastrophic forgetting. This issue is further aggravated in CSS where, at each step, old classes from previous iterations are collapsed into the background. In this paper, we propose Local POD, a multi-scale pooling distillation scheme that preserves long- and short-range spatial relationships at feature level. Furthermore, we design an entropy-based pseudo-labelling of the background w.r.t. classes predicted by the old model to deal with background shift and avoid catastrophic forgetting of the old classes. Our approach, called PLOP, significantly outperforms state-of-the-art methods in existing CSS scenarios, as well as in newly proposed challenging benchmarks." @default.
• W3110282355 created "2020-12-07" @default.
• W3110282355 creator A5022871131 @default.
• W3110282355 creator A5037619048 @default.
• W3110282355 creator A5052337951 @default.
• W3110282355 creator A5082314947 @default.
• W3110282355 date "2021-06-01" @default.
• W3110282355 modified "2023-09-25" @default.
• W3110282355 title "PLOP: Learning without Forgetting for Continual Semantic Segmentation" @default.
• W3110282355 cites W1745334888 @default.
• W3110282355 cites W1821462560 @default.
• W3110282355 cites W1901129140 @default.
• W3110282355 cites W1903029394 @default.
• W3110282355 cites W2037227137 @default.
• W3110282355 cites W2060277733 @default.
• W3110282355 cites W2108598243 @default.
• W3110282355 cites W2116522068 @default.
• W3110282355 cites W2126204609 @default.
• W3110282355 cites W2134270519 @default.
• W3110282355 cites W2179352600 @default.
• W3110282355 cites W2194775991 @default.
• W3110282355 cites W2340897893 @default.
• W3110282355 cites W2412782625 @default.
• W3110282355 cites W2560023338 @default.
• W3110282355 cites W2560647685 @default.
• W3110282355 cites W2561196672 @default.
• W3110282355 cites W2561238782 @default.
• W3110282355 cites W2583761661 @default.
• W3110282355 cites W2630837129 @default.
• W3110282355 cites W2737258237 @default.
• W3110282355 cites W2737492962 @default.
• W3110282355 cites W2771964490 @default.
• W3110282355 cites W2786446225 @default.
• W3110282355 cites W2807445254 @default.
• W3110282355 cites W2884282566 @default.
• W3110282355 cites W2890782586 @default.
• W3110282355 cites W2895281799 @default.
• W3110282355 cites W2899771611 @default.
• W3110282355 cites W2902625698 @default.
• W3110282355 cites W2913699173 @default.
• W3110282355 cites W2922466325 @default.
• W3110282355 cites W2925387673 @default.
• W3110282355 cites W2932484476 @default.
• W3110282355 cites W2947461406 @default.
• W3110282355 cites W2948734064 @default.
• W3110282355 cites W2954929116 @default.
• W3110282355 cites W2955058313 @default.
• W3110282355 cites W2962724315 @default.
• W3110282355 cites W2963073217 @default.
• W3110282355 cites W2963540014 @default.
• W3110282355 cites W2963542991 @default.
• W3110282355 cites W2963545832 @default.
• W3110282355 cites W2963559848 @default.
• W3110282355 cites W2963560049 @default.
• W3110282355 cites W2963588172 @default.
• W3110282355 cites W2963727650 @default.
• W3110282355 cites W2963813662 @default.
• W3110282355 cites W2963881378 @default.
• W3110282355 cites W2964118293 @default.
• W3110282355 cites W2964151081 @default.
• W3110282355 cites W2964189064 @default.
• W3110282355 cites W2964309882 @default.
• W3110282355 cites W2971113548 @default.
• W3110282355 cites W2971234316 @default.
• W3110282355 cites W2972285644 @default.
• W3110282355 cites W2972313371 @default.
• W3110282355 cites W2981689412 @default.
• W3110282355 cites W2981864462 @default.
• W3110282355 cites W2994827984 @default.
• W3110282355 cites W3011986058 @default.
• W3110282355 cites W3021931813 @default.
• W3110282355 cites W3024202154 @default.
• W3110282355 cites W3028392891 @default.
• W3110282355 cites W3034435444 @default.
• W3110282355 cites W3034502973 @default.
• W3110282355 cites W3034679848 @default.
• W3110282355 cites W3034856281 @default.
• W3110282355 cites W3037416150 @default.
• W3110282355 cites W3037765293 @default.
• W3110282355 cites W3040304705 @default.
• W3110282355 cites W3043569052 @default.
• W3110282355 cites W3049435445 @default.
• W3110282355 cites W3094953068 @default.
• W3110282355 cites W3100156920 @default.
• W3110282355 cites W3101570898 @default.
• W3110282355 cites W3107810305 @default.
• W3110282355 cites W3108280718 @default.
• W3110282355 cites W3109301572 @default.
• W3110282355 hasPublicationYear "2021" @default.
• W3110282355 type Work @default.
• W3110282355 sameAs 3110282355 @default.
• W3110282355 citedByCount "2" @default.
• W3110282355 countsByYear W31102823552021 @default.
• W3110282355 crossrefType "proceedings-article" @default.
• W3110282355 hasAuthorship W3110282355A5022871131 @default.
• W3110282355 hasAuthorship W3110282355A5037619048 @default.
• W3110282355 hasAuthorship W3110282355A5052337951 @default.
• W3110282355 hasAuthorship W3110282355A5082314947 @default.
• W3110282355 hasBestOaLocation W31102823551 @default.
• W3110282355 hasConcept C108583219 @default.

• next